Manipulating Political Stock Markets (Working Paper)
Total Page:16
File Type:pdf, Size:1020Kb
Manipulating Political Stock Markets: A Field Experiment and a Century of Observational Data* Paul W. Rhode** Koleman S. Strumpf Univ. of Arizona and NBER Univ. of Kansas School of Business June 2008 Preliminary Abstract Political stock markets have a long history in the United States. Organized prediction markets for Presidential elections have operated on Wall Street (1880-1944), the Iowa Electronic Market (1988-present), and the internet (2000-present). Proponents claim such markets efficiently aggregate information and provide forecasts superior to polls. An important counterclaim is that such markets may be subject to manipulation by interested parties. We investigate the impact of actual and alleged speculative attacks— large trades, uninformed by fundamentals, intended to change prices— in political stock markets. First we report the results of a field experiment involving a series of planned, random investments-- accounting for two percent of total market volume-- in the Iowa Electronic Market in 2000. We next examine the historical Wall Street markets where political operatives from the contending parties actively and openly bet on city, state and national races; the record is rife with accusations that parties tried to boost their candidates through investments and wash bets. Finally, we investigate the speculative attacks on TradeSports market in 2004 when a single trader made a series of large investments in an apparent attempt to make one candidate appear stronger. In the cases studied, the speculative attack initially moved prices, but these changes were quickly undone and prices returned close to their previous levels. We find little evidence that political stock markets can be systematically manipulated beyond short time periods. Our results potentially have implications for trader behavior in broader financial markets. *We thank Tim Groseclose, Robin Hanson, Donald Luskin, Tom Mroz, Sergio Parreiras, Marco Ottaviani, Charles Plott, David Primo, Mark Stegeman, Justin Wolfers, participants at the DIMACS Workshop on Markets as Predictive Devices, the 2005 International Symposium on Forecasting, London Business School Information and Prediction Markets Conference, 2006 NBER DAE Program Meeting, 2006 Prediction Markets Summit East, and workshops at Arizona, Iowa State, UNC Chapel Hill, UNC Greensboro, University of Virginia, Yale, and WZB for assistance and comments. We are particularly grateful to Intrade (part of the Trade Exchange Network) and Betfair for providing data from the 2004 Presidential Election market, Steve Cornwell of Intertops.com and Ken Kittlitz of the Foresight Exchange for data on the 2000 Presidential Election market, and Bo Cowgill of Google for detailing that company’s internal prediction markets. **Corresponding author. McClelland Hall Room 401NN. Department of Economics. Univ. of Arizona. Tucson, AZ 85718. email: [email protected] I. Introduction Prediction markets involve contracts which have payoffs explicitly linked to future events. An example is a binary option which pays a dollar on the outcome of a specific event, such as a candidate’s victory or an on-time product launch. An efficient prediction market aggregates available information, yielding prices that are the best forecast of the event’s probability. Such forecasts can be socially valuable if the market involves a topic of great importance. Prediction markets are currently the subject of intensive research in fields ranging from economics to political science to computer science (Berg, et al, 2005; Hanson, 1999; Pennock, 2004; Wolfers and Zitzewitz, 2004; Ledyard, 2005; Snowberg, Wolfers and Zitzewitz, 2007). There is also growing interest outside of academia. In the popular press, James Surowiecki (2004) has championed the Wisdom of Crowds and in the private sector, Abbott Labs, Best Buy, Corning, Electronic Arts, General Electric, Goldman Sachs, Google, Hewlett-Packard, Intel, Lilly, Microsoft, Siemens, and Yahoo! have set up internal prediction markets. Public prediction markets on current events, economic outcomes, and even the weather have become increasing common as the internet has opened access to a growing number of on-line sites such as TradeSports-Intrade, Betfair, and News Futures. In political prediction markets alone, one-hundred thousand participants conducted over three million trades in 2006. The hope is such markets can aid forecasting and improve decision-making in economic policy, corporate project selection, and other areas. Skeptics have advanced several theoretical challenges to the efficiency and predictive power of these markets. For example Manski (2006) questions the received wisdom that prices can be interpreted as probabilities. In his model, market prices only provide information about the wide interval in which mean beliefs over probabilities lie. Responses include Wolfers and Zitzewitz (2005), Gjerstad (2005), and Ottaviani and Sørensen (2005). A more damaging challenge to the forecasting ability of prediction markets is the possibility that investors could distort prices away from fundamentals for the strategic purpose of influencing the expectations and actions of others. Such manipulation is an inherent danger for prediction markets for several reasons. First, the potential reward 1 from a successful manipulation can far exceed the financial resources needed to implement it. Consider the case of political markets. The reported odds might influence the election if voters are unwilling to support a candidate who is faring poorly. Such changes in price require relatively small stakes (even big political markets attract volume in the tens of millions of dollars) but can shape a very large outcome (the federal budget involves trillions of dollars). As their visibly rises, prediction markets will likely become increasingly tempting targets for manipulators. A second issue is that insiders are not prohibited from trading in these markets. On-line services such as TradeSports or the Hollywood Stock Exchange do not explicitly ban insider traders and the Commodities Futures Trading Commission (CFTC) regulations permit such trading in future markets. So traders cannot be sure a seemingly inexplicable price change is the result of a manipulator or rather is the result of some new private information. Theory aside, the potential for manipulation is often levied as a criticism of prediction markets. Stiglitz (2003) criticized the proposed Policy Analysis Market, a heavily publicized futures market on Middle East economic and military events, because it “could be subject to manipulation.” Such behavior is also an issue for internal corporate markets, with the organizer of Google’s market noting that he “repeatedly encountered concern about the potential for market manipulation from average employees as well as senior leadership” (Cowgill, 2006). At least one trader at Google later admitted to attempting to manipulate prices. Manipulation attempts are readily observed in the field. As a motivating example, consider movements in the price of shares for George W. Bush in the 2004 US Presidential election at TradeSports.1 Figure 16 displays the price and volume during September and October. Shortly after 2:30 pm (EDT) on Friday, October 15, 2004, the TradeSports odds price on the re-election of President Bush began to fall precipitously. 1This was a large and influential market, attracting more than $15M in trade volume. Shares in the main election market paid a fixed amount if Bush won, and the prices were scaled between zero and a hundred to give the usual probability interpretation. TradeSports markets are listed at http://www.tradesports.com. It is part of the Trade Exchange Network which provides an electronic matching service for trading futures on sports, entertainment, legal, and political events. The company, based in Dublin, Ireland, was founded in 2001. Its shares pay $10 upon winning but are quoted between 0 and 100. When share prices are between 6 and 94, or exactly 0 or 100, then TradeSports charges a commission of 0.04 dollars (about 0.8 percent) per shared trades. Outside that range to the extremes the commission rate is 0.02. 2 From a plateau of 54 points at 2:30 pm, a series of thirty trades in less than a second dropped the price to 48 at 2:31 pm. After stabilizing for two minutes, another rapid set of trades led prices to tumble to 10 at 2:33 pm. Thus prices fell by 44 points in just three minutes, suggesting that Bush went from a slight favorite to serious underdog. This sharp drop was the most dramatic of a series of trades that National Review Online blogger Donald Luskin soon charged were politically-motivated speculative attacks on Bush futures “to sway the election towards Kerry.”2 Reports circulated that George Soros was behind the October 15 plunge as well as earlier bear raids on Bush. Such rumors gained currency when a TradeSports press release, publicized in Wall Street Journal and Time, confirmed that the large trades of a single investor produced the October 15 price moves.3 The press release asserted “Bush contract has become the battle ground of wills between a cadre of large, well financed rogue traders seemingly bent on driving down the Bush re-election contract and a growing list of financial traders who think they can predict the outcome of this election.” In addition to the October 15 episode, the price of the “Bush Winner” contract also experienced a a 13-point drop during a fourteen minute period around 12 pm EDT on Monday, September 13. Figure 17 shows